references/coding-interview-rubric.md

# Coding interview rubric

Evaluate observable evidence:
- Correctness: actual behavior vs specification, including empty and boundary cases.
- Approach: reasonable algorithm/data structures, invariant correctness.
- Complexity: justified time and space relative to constraints.
- Testing: representative examples, edge cases, invalid inputs where relevant.
- Code quality: readability and clear failure handling.
- Communication: assumptions, explaining trade-offs and correction.

When coding execution is available, test with clean fixtures and report actual commands/results. Otherwise mark 'reasoned, not executed'. Avoid misleading single-score hire/no-hire declarations; use qualitative coaching or explicit rubric weights only if requested.

Feedback template: observed evidence -> exact flaw -> corrected idea -> minimum reproducible test -> practice target.